AI Biodiversity Monitoring Developer

Zelux Studio

Remote · $80k – $100k / year · Full-time · Posted 4 months ago

This role is no longer accepting applications

Don't miss the next role like this

Create a free profile with your skills. We match you to new AI roles as they open, and companies hiring can contact you directly.

Create free profile, 2 minutes

Similar open roles you can apply to now:

See all open AI jobs

This role is no longer accepting applications

Don't miss the next role like this

Create a free profile with your skills. We match you to new AI roles as they open, and companies hiring can contact you directly.

Create free profile, 2 minutes

Similar open roles you can apply to now:

See all open AI jobs

About the role

Identifying species from camera trap footage used to require a trained ecologist reviewing every frame. With 14,000 camera traps deployed across 6 conservation projects, that's 2 million images per month — physically impossible to review manually. Our computer vision pipeline classifies 94% of images automatically, freeing ecologists to focus on the ambiguous 6% and the actual conservation decisions. The 6% that fails automation includes low-light night images, heavily obscured animals, species pairs that differ by subtle markings, and novel species not in the training set. You'll improve model performance on these failure modes using targeted data collection, active learning, and few-shot classification techniques that work with the small labeled datasets available for rare species. This is a meaningful role in conservation technology. Our clients include WWF, Wildlife Conservation Society, and several national park authorities across Africa and Southeast Asia.

Fully remote.

Responsibilities

Improve classification accuracy on failure modes: low-light, occlusion, rare species
Build active learning pipelines to efficiently prioritize ambiguous images for review
Implement few-shot classification for rare species with limited training data
Design out-of-distribution detection for novel species not in the training set
Create evaluation frameworks using ecologist-validated ground truth
Collaborate with field teams on data collection strategies for underrepresented species

Requirements

3+ years in computer vision engineering with classification or detection experience
Experience with active learning, few-shot learning, or out-of-distribution detection
Strong PyTorch skills and familiarity with fine-tuning pre-trained vision models
Ability to work with small, imbalanced labeled datasets
Genuine interest in conservation technology or environmental monitoring